Your screenshot is yours.
Your screenshots are analyzed solely to generate your reading. They are not used for AI training, advertising, or sold to third parties. Screenshots are automatically deleted after processing whenever possible.
What we promise
- Screenshots are used only to generate your reading.
- Screenshots are never used to train AI models.
- Screenshots are never sold, shared, or used for advertising.
- Screenshots are not added to any permanent archive.
- Screenshots are not stored in databases, storage buckets, or long-lived logs.
How a reading actually works today
When you upload a screenshot, your browser reads the image into memory and posts it directly to our reading function. The image is forwarded once to the Lovable AI Gateway (which routes the request to a vision model, currently Google Gemini) and the model returns a short JSON reading. That JSON — not the image — is what you see on the reading page.
The reading JSON is held in your browser's sessionStorageso the reading page can display it after you navigate. It is cleared automatically when you close the tab.
Where the screenshot lives
- Where it is stored
- Nowhere persistent. The image exists only in memory during the single request that generates your reading.
- How long it is kept
- For the duration of the request — typically a few seconds. There is no database row, no storage bucket, and no file on disk.
- Who can access it
- The reading server function and the upstream AI provider (Lovable AI Gateway → Google Gemini) for the single API call. No human at Bad Thoughts Studio sees your screenshot.
- Logs
- We do not log image contents. If a request fails, only a short error message (no image data) is captured.
- AI training
- The upstream model is called via an API and the request is not used to train models. We never share your screenshot with advertisers or third parties.
Deep Reading (paid, future)
If a paid Deep Reading workflow ever requires temporary storage to complete the order, we will retain only the minimum data needed to finish that one reading and delete the screenshot automatically as soon as the reading is delivered. No permanent archive will be kept.
What changed
We audited the current implementation and confirmed there is no database table, storage bucket, or server-side cache holding uploaded screenshots. The upload flow already passes images in-memory only. We added this page and an in-product privacy notice on the reading screen so the behavior is documented and visible.
Questions? hello@badthoughts.studio